Summary Ralph Wiggum is an autonomous AI coding system enabling spec-driven development. It allows AI agents to iteratively implement, verify, and commit code based on predefined specifications. This tool is ideal for developers aiming to automate coding workflows and accelerate development by leveraging AI within a structured, verifiable process.

How It Works The system combines an iterative bash loop with SpecKit-style specifications. Each cycle focuses on a single task: implementing code, running tests, and committing changes. The AI signals completion (<promise>DONE</promise>) only when all acceptance criteria are met and verified. Key principles include "Fresh Context Each Loop" for isolation and "Shared State on Disk" (IMPLEMENTATION_PLAN.md) for continuity, ensuring rigorous adherence to requirements.

Quick Start & Requirements Install via Agent Skill: npx add-skill fstandhartinger/ralph-wiggum or openskills install fstandhartinger/ralph-wiggum. Full setup involves INSTALLATION.md (guided interview ~3-5 min, or manual). Requires an AI agent capable of executing shell commands (bash/PowerShell) and interacting with a codebase. Quick setup is ~1 minute.

Highlighted Details

• Iterative Self-Correction: Each loop handles one task, implements, verifies, and commits.
• Spec-Driven Development: Driven by detailed specifications with clear, testable acceptance criteria.
• Completion Verification: Agent outputs <promise>DONE</promise> only after criteria verification and commits.
• Context Management: Uses "Fresh Context Each Loop" and persists state via IMPLEMENTATION_PLAN.md.
• Multi-AI Support: Integrates with Claude Code, Codex, Gemini, Copilot.
• Comprehensive Logging: Detailed logs for session and iteration outputs.
• Stuck Spec Detection: Flags specs attempted >10 times without completion.

Maintenance & Community The README does not detail specific contributors, community channels (Discord/Slack), or a public roadmap.

Licensing & Compatibility Released under the MIT License, permitting broad use, modification, and distribution, including for commercial purposes. No specific compatibility restrictions are noted.

Limitations & Caveats The project carries a "Use at your own risk" disclaimer, recommending isolated environments due to AI agent autonomy. Effectiveness hinges on high-quality, testable specifications. A mechanism flags "stuck" specs, requiring manual intervention or splitting tasks.